CodeTools proposal: "friday stats"

[Jonathan Gibbons]

On 07/03/2013 03:45 AM, Aleksey Shipilev wrote:
> On 07/03/2013 02:09 PM, Jonathan Gibbons wrote:
>> On 07/01/2013 07:24 AM, Dalibor Topic wrote:
>>> On 6/18/13 8:43 AM, Aleksey Shipilev wrote:
>>>> +1.
>>>>
>>>> On the related note, do we want to set up the Sonar [1] instance to
>>>> analyze OpenJDK codebase? In my experience, _that_ really highlights the
>>>> problematic areas.
>>> See
>>> http://mail.openjdk.java.net/pipermail/discuss/2011-November/002204.html
>>>
>>> cheers,
>>> dalibor topic
>> The main issue is getting manageable reports that people can handle.
>> There's no point generating a report that shows we have ten gazillion
>> issues to fix, because there will be no significant reward for anyone
>> working to reduce the numbers. You need to be able to target specific
>> areas for which there is a hope of making progress.
> Sure. BTW, I asked Evgeny for the JDK 8 instance, where the development
> is going on:
>     https://twitter.com/shipilev/status/351719579724161026
>
> -Aleksey.
>

I scanned through some of this report. It is hard to know how we can 
best make use of it. In addition, some of the "major issues" don't seem 
so major to me -- like "superfluous parentheses".  And "security issues" 
come up as just "info".

I don't think one report for the JDK is the best way to go. If we could 
configure a report for each of the major component areas, and configure 
the messages that are generated, then we might start to get somewhere.

I see some of the issues are close to "code style" issues. These can 
sometimes be very difficult to fix, because of the risk of making it 
hard to integrate changes in progress in other repos, and making it 
harder to backport changes to other releases.

-- Jon